R ecent controlled studies demonstrate that aversion therapies can be effective treatments for certain behaviour disorders such as alcoholism (Ashem and Donner, 1968) and sexual deviation (Feldman and MacCulloch, 1971). However, the mechanism of action in the various forms of aversion therapy has not been determined. As one form of aversion therapy, covert sensitization employs descriptions of noxious scenes as aversive stimuli. An earlier study (Barlow, Leitenberg and Agras, 1969) suggested that pairing descriptions of noxious scenes with scenes of the undesired behavior was the critical procedure in covert sensitization. There, a single case design was used with two sexual deviates. Sexually arousing scenes were paired with descriptions of nausea and vomiting during the first phase, while in a second phase the same sexually arousing scenes were presented alone, and finally, pairing was reinstated. Put another way, these phases represent acquisition, extinction and re-acquisition respectively. During the pairing of scenes in the first phase, deviant arousal declined, rose again when the pairing of scenes was removed, and dropped in the final phase when pairing was reinstated. This appeared to demonstrate the effectiveness of pairing in the procedure. It is possible, however, that therapeutic instructions and the resulting expectancy of improvement present in both covert sensitization phases, were responsible for the effectiveness of the treatment, since the patient may have viewed the middle extinction phase as non- therapeutic. To test this notion, the procedure of the first experiment was changed somewhat. Four homosexual Ss were told that the acquisition procedure (covert sensitization) with pairing would temporarily worsen their sexual deviation and that the extinction procedure (no pairing) was therapeutic. The experiment began with the no pairing procedure with therapeutic instructions to create the desired expectancies at the outset and to test for an initial placebo effect. In addition, an objective measure of sexual arousal, penile circumference change, was recorded. If pairing is the necessary and sufficient procedure in covert sensitization, then deviant sexual arousal should decrease during covert sensitization despite instructions to the contrary and, conversely, no improvement should occur during the no pairing, placebo phase.